In vivo speciation studies and antioxidant properties of bromine in Laminaria digitata reinforce the significance of iodine accumulation for kelps.

The metabolism of bromine in marine brown algae remains poorly understood. This contrasts with the recent finding that the accumulation of iodide in the brown alga Laminaria serves the provision of an inorganic antioxidant - the first case documented from a living system. The aim of this study was to use an interdisciplinary array of techniques to study the chemical speciation, transformation, and function of bromine in Laminaria and to investigate the link between bromine and iodine metabolism, in particular in the antioxidant context. First, bromine and iodine levels in different Laminaria tissues were compared by inductively coupled plasma MS. Using in vivo X-ray absorption spectroscopy, it was found that, similarly to iodine, bromine is predominantly present in this alga in the form of bromide, albeit at lower concentrations, and that it shows similar behaviour upon oxidative stress. However, from a thermodynamic and kinetic standpoint, supported by in vitro and reconstituted in vivo assays, bromide is less suitable than iodide as an antioxidant against most reactive oxygen species except superoxide, possibly explaining why kelps prefer to accumulate iodide. This constitutes the first-ever study exploring the potential antioxidant function of bromide in a living system and other potential physiological roles. Given the tissue-specific differences observed in the content and speciation of bromine, it is concluded that the bromide uptake mechanism is different from the vanadium iodoperoxidase-mediated uptake of iodide in L. digitata and that its function is likely to be complementary to the iodide antioxidant system for detoxifying superoxide.

Introduction

Two centuries ago, the elements bromine and iodine were discovered in seawater and seaweed ashes, respectively (Balard, 1826; Gay-Lussac, 1813). Due to their unique evolutionary history and phylogenetic distance from other important eukaryotic lineages (Bhattacharya ; Baldauf, 2003), brown algae present some remarkable chemical and physiological adaptations which are also reflected at the genome level (Cock ), making them fascinating experimental models, not only for phycologists, but for a community of interdisciplinary researchers. This includes the recently described function of iodide as an extracellular antioxidant protecting the surface of Laminaria digitata (Hudson) Lamouroux against oxidative stress (Küpper ). In fact, this constituted the first documented case of an inorganic antioxidant in a living system – and the chemically simplest antioxidant known.

Indeed, brown algae of the genus Laminaria are the strongest accumulators of iodine in life on Earth (Saenko ; Küpper ; Ar Gall ), and they are a major contributor to the biogeochemical flux of iodine (McFiggans ) and, to a lesser extent, brominated and iodinated halocarbons to the atmosphere (Carpenter ). However, while the uptake, metabolism, and biogeochemical cycling of iodine by Laminaria are well studied (Küpper ), hardly anything is known about bromine in this context. The uptake of iodide from seawater in Laminaria involves vanadium haloperoxidases (VHPOs) (Küpper ) and its strongest accumulation in this species seems to be linked to the presence of a particular VHPO subclass, the iodoperoxidases, specific for iodide oxidation (Colin , 2005). Most of the iodine is accumulated in the apoplast of the cortical cell layer (Verhaeghe ). Following a hypothesis from the biomedical field about a potential ancestral role of iodide as an antioxidant in the evolution of the thyroid in vertebrates (Venturi and Venturi, 1999), it was shown that iodide is the accumulated form of iodine in Laminaria, and that it indeed serves as a simple, inorganic antioxidant, protecting the apoplast and thallus surface against both aqueous and gaseous oxidants (Küpper ). This study also revealed that the iodide in the liquid microlayer on the emerged Laminaria thallus surface can efficiently scavenge ozone from the gas phase, resulting in atmospheric particle formation, which can act as cloud condensation nuclei (McFiggans ; Palmer ). In this study, X-ray absorption spectroscopy (XAS) has proven to be a suitable, non-invasive tool to probe the chemical state and solution environment of accumulated iodine. However, this technique is equally suitable to investigate the speciation of bromine in kelps (Feiters ,b; Strange and Feiters, 2008).

The first line of defence against pathogens in Laminaria is an oxidative burst (Küpper ), which is considered a central element of eukaryotic defence in general (Wojtaszek, 1997) and which, in the case of kelp species, serves to control bacterial biofilms (Küpper ). Among its triggers are oligoguluronates (GG) (Küpper , 2002), bacterial lipopolysaccharides (Küpper ), prostaglandin A2 (Zambounis ), methyl jasmonate, and polyunsaturated free fatty acids (Küpper ). In L. digitata, early transcriptional defence responses are similar to those in land plants but also involve tightly regulated halogen metabolism which might play roles in more sophisticated chemical defence reactions including distance signalling (Cosse ; Thomas ). In brown algae, bromide and VHPOs have also been shown to catalyse oxidative cross-linking between cell-wall polymers, suggesting a role in spore and gamete adhesion and cell-wall strengthening (for reviews, see Potin and Leblanc, 2006; La Barre ).

This study addresses the biological significance of bromine versus iodine metabolism in Laminaria. First, iodine and bromine levels in different types of tissues of Laminaria were investigated using inductively coupled plasma (ICP) MS. Then K-edge XAS was applied to probe the stored chemical state of these elements under different physiological conditions in vivo (both steady-state unstressed and oxidatively stressed) in the different tissue types of Laminaria and to draw comparisons and potential functional links between bromine and iodine metabolism. Finally, the potential antioxidant properties of bromide and iodide were compared in whole-blood oxidative burst assays and in vitro electron paramagnetic resonance (EPR) assays.

Materials and methods

Oxidative stress experiments for XAS

Oxidative stress experiments for XAS were conducted as described previously (Küpper ). L. digitata sporophytes (approx. 2–100cm in size) were collected in the sublittoral at Roscoff (Brittany) and Helgoland (Germany) and kept in tanks with aerated, running seawater between 4–10 °C until use. Oligoguluronate elicitors (GG) were applied at a final concentration of 100 μg ml–1. Typically, experiments were conducted with around 5g Laminaria fresh weight in 100ml natural seawater, from which both tissue samples of approximately 0.5g and aliquots of seawater medium of 15ml were removed directly at the onset of the stress, and then at intervals of 1, 3, 5, 10, 15, 30, 45, 60, 180 minutes, and 24 hours, respectively. L. digitata tissue samples were cut out of the phylloid blade and blotted dry on paper tissue. They were immediately fitted into a plexiglass frame and sealed with Capton tape. All samples were immediately frozen in liquid nitrogen. This experiment was repeated around 15 times with minor variations in sample volume, total algal biomass, and seawater volume.

Preparation of different Laminaria tissue samples, extracts, and enzymes for XAS

Isolated cell walls and a polyphenol-enriched fraction were prepared from L. digitata sporophytes collected in the sublittoral at Roscoff (Brittany), as described previously (Mabeau and Kloareg, 1987; Connan ). The polyphenol-enriched fraction is a methanol/water extract (50:50) of Laminaria blades at 40 °C followed by vacuum drying. An industrial alginate sample was obtained from a major alginate manufacturer (DANISCO); the production followed a well-established technique (McHugh, 1987). The other tissue samples were prepared as described before (Küpper ), by dissecting, blotting/drying, and sealing in a plexiglass frame sealed with Capton tape, followed by flash-freezing in liquid N2. The extraction and purification of the native VHPOs from L. digitata have been described previously (Colin ).

ICP-MS analysis of total iodine bromine

Iodine and bromine contents in freeze-dried L. digitata tissue and cell-wall samples were analysed by ICP-MS after pyrohydrolysis separation, as described previously (Schnetger and Muramatsu, 1996; Chai and Muramatsu, 2007). Concentrations obtained are given in ppm (i.e. μg halogen per g of freeze-dried material).

X-ray absorption spectroscopy

Bromine and iodine K-edge XAS measurements and extended X-ray absorption fine structure (EXAFS) data reduction were carried out at the EMBL Outstation Hamburg at DESY, Germany, as described previously (Feiters ). PDB files for the multiple scattering units of the aromatic amino acids were generated using ChemBio3D Ultra.

Neutrophil oxidative burst assay for determining the antioxidant potential of bromide

Blood was withdrawn from healthy volunteers into blood collection tubes containing sodium citrate (Sarstedt). Neutrophils were isolated using density gradient centrifugation and resuspended in Hanks Balanced Salt solution at a density of 1,000,000 cells ml–1. Neutrophils were mixed 1:1 with 20 μM 2′,7′-dichlorofluorescein diacetate and incubated (37 °C, 5% CO2) for 30 minutes prior to seeding into a 96-well plate. Test compounds were added to triplicate wells at a range of concentrations (0.01–100mM). Oxidative burst was stimulated by addition of 0.1nM phorbol myristate acetate (Repine ). The plate was incubated at 37 °C with fluorescence readings taken every 10 minutes. Vmax was calculated for each test group over four data points. Percentage activity was calculated by comparing test group Vmax to control Vmax.

EPR antioxidant assay

The antioxidant potential of KBr and KI was determined by EPR. Stock solutions of KI and KBr (0.10M in H2O) were diluted as required to the final working concentration. Both compounds were tested with the radical-generating systems menadione (in DMSO) and pyrogallol (in H2O) at a final concentration of 150 μM in the presence of the spin trap, tempone-H (50 μM). Both menadione and pyrogallol undergo auto-oxidation under oxic (atmospheric oxygen) conditions to generate oxygen-centred radicals. The spin-trap chosen is recognized to be selective for oxygen-centred radicals (Meja ). Menadione incubation was performed in double-distilled H2O, and pyrogallol incubation in PBS, with appropriate vehicle controls (i.e. PBS or H2O only) run as blanks. Incubations were performed at 37 °C with readings taken at 0, 15, 30, 45, and 60min using a benchtop MS200 X-Band EPR spectrometer (Magnettech, Berlin, Germany) set with the following parameters: microwave frequency 9.30–9.55 GHz; B0 field 3344 Gauss; sweep width 50 Gauss; sweep time 60 s; modulation amplitude 2.0 Gauss; microwave power 10 mW. Formation of the spin-adduct (4-oxo-tempo) by oxidizing radical species was measured and the amplitude of the first derivative spectra obtained (arbitrary units) was plotted against time following subtraction of appropriate blank measurements. Radical generation was estimated by measuring the area under the resultant linear plot for each individual 60-min incubation, prior to mean and standard error calculations (Graphpad Prism version 5.0).

Results and discussion

Very little is known about the chemical biology and atmospheric chemistry of bromine, despite the accumulating wealth of knowledge for iodine regarding physiology and biochemistry of animal and algal systems and its importance in biogeochemical cycles (Carpenter ; Küpper ). Brominating oxidants are formed by eosinophils, capable of destroying a wide range of prokaryotic and eukaryotic targets (Weiss ) and marine macroalgae are known to emit bromocarbons at high rates (Carpenter and Liss, 2000; Carpenter ). The recent discovery that the accumulation of iodide serves the provision of a simple, inorganic antioxidant in the kelp Laminaria (Küpper ) prompted the current investigation of the chemical speciation and biological significance of bromine in this important model system with a similar, interdisciplinary array of approaches and techniques.

The first objective was to determine bromine levels in various types of Laminaria tissues by ICP-MS. Most of the samples studied had previously served for XAS analyses reported below. Overall, the ICP-MS results of bromine levels in L. digitata blades and stipes (Table 1; for Laminaria anatomy, see Supplementary Fig. S1, available at JXB online) were consistent with previous results measured in freeze-dried and chemically fixed tissues using neutron activation analysis (Verhaeghe ). The highest bromine levels were encountered in the holdfast (≈ 2000 ppm) and cortical stipe tissues (≈ 1800–1900 ppm), while the lowest concentrations were found in foliar blade (966±77 ppm) and medullary stipe tissues (654±26 ppm). The highest iodine levels (also measured by ICP-MS), up to more than 57,000 ppm, were observed in outer cortical tissues of Laminaria stipes, followed by the inner cortex (≈40,000 ppm) while the lowest tissue concentrations were found in the medulla of stipes (≈1000 ppm). Foliar blade tissues contained close to 5000 ppm (but due to the anatomy of the blade, it was not possible to separate cortical from medullary tissue; it is assumed that the cortex would show much higher iodine levels than the medulla). There was a clear pattern of higher bromine and iodine concentrations in cortical, foliar blade and holdfast tissues as opposed to the thallus medulla, underlining the characteristics of halogen accumulation as a brown algal surface process. Overall, accumulation of iodine was much more pronounced than that of bromine with total concentrations up to over an order of magnitude higher. Considering the much lower concentration of iodine in seawater (0.5 μM in total, with about equal levels of iodide and iodate) in comparison to bromide (0.8mM), the accumulation factor of iodine from seawater to Laminaria tissues (around five orders of magnitude) was much more extreme than for bromine (less than 1 order of magnitude). In contrast, industrial sodium alginate from L. digitata contained very low and similar bromine (12±4 ppm) and iodine levels (20±3 ppm).

Table 1.

Total non-volatile iodine and bromine content of samples of various tissue parts and cell-wall components of Laminaria digitata. Determined by ICP-MS. Algal tissue samples were freeze-dried and ground in liquid nitrogen. n, technical replicates.

Sample	Iodine concentration (mean ppm)	n	Standard deviation (%)	Bromine concentration (mean ppm)	n	Standard deviation (%)
Foliar blade tissues from Laminaria digitata
1st whole blade	15505	2	16	714	1	–
2nd whole blade	4834	4	1	966	4	8
Stipe tissues from Laminaria digitata
Whole stipe sections (about 0.4cm in diameter)	12790	2	6	1465	1	–
1st dissected medullary tissues	1188	4	6	1019	4	6
2nd dissected medullary tissues	1670	3	5	654	3	4
Dissected inner cortex tissues	39708	4	6	1940	2	1
Dissected outer cortex and epidermis tissues	57717	4	16	1794	1
Holdfast tissues from Laminaria digitata
1st whole holdfast	8057	4	22	2016	4	14
2nd whole holdfast	9715	4	3	1864	4	9
Industrial sodium alginate batch extracted from Laminaria digitata (Danisco, Landerneau, France)	20	3	15	12	3	37

XAS was used to elucidate the in vivo speciation of bromine and to compare it to that of iodine. Representative Br K-edge X-ray absorption near-edge spectra are shown in Supplementary Fig. S2 and Br K-edge extended X-ray absorption fine structure (EXAFS) and corresponding Fourier transforms are shown with their simulations in Figs. 1–3. As described in detail in the supporting information (available at JXB online), three types of contributions to the biological Br EXAFS spectra obtained with Laminaria thallus samples could be distinguished, which were representative of three different chemical environments, viz. Br– surrounded by hydrogen-bonding molecules (Fig. 1), Br covalently incorporated into aromatic moieties (Fig. 2), and Br– incorporated in solid KCl (Fig. 3), with examples of the corresponding edge X-ray absorption near-edge spectra in Supplementary Fig. S2 (traces A, C/D, and B, respectively). Bromine and iodine K-edge EXAFS and corresponding Fourier transforms are given in Supplementary Figs. S3 and S4, and all iterative refinement parameters of the simulations for Br and I are given in Supplementary Tables S1 and S2, respectively. The XAS results in Fig. 1 clearly show that bromide is the dominant form of bromine in Laminaria, and that it is present in a non-covalently bound form, associated with heteroatoms from biomolecules by hydrogen-bonding. Br was found to be predominantly hydrogen-bonded to biomolecules in fresh and lyophilized Laminaria (Fig. 1, traces D and E), but the larger number of oxygen atoms found for oligo-GG stress (Fig. 1, trace C) compared to fresh Laminaria indicated that a larger fraction of it was hydrated under oligo-GG stress rather than bound to biomolecules in fresh Laminaria. This is analogous to the situation reported earlier (Küpper ) for iodide in Laminaria, and it is in line with the differences observed between the dissolving of iodide by small solvent molecules, resulting in interaction with many solvent atoms, or by larger solvent molecules, where fewer solvent atoms can interact (Tanida and Watanabe, 2000). In contrast, the XAS spectra of freeze-dried Laminaria tissues rehydrated with 2mM H2O2 (Supplementary Fig. S3, trace A) could be interpreted by the incorporation of bromide in aromatic substrates both in the rehydrated thallus and in the cell-wall fraction, either directly by the oxidative activation of Br– by H2O2 or by the action of haloperoxidases. An analogous observation was made for iodide in Laminaria (Supplementary Fig. S4, trace H; Küpper ). Of the Laminaria parts, the cell wall was the only one that displayed this spectrum (Supplementary Fig. S3, trace B). While the spectrum of the Laminaria holdfast (Fig. 1, trace F) resembled that of whole Laminaria, indicating the presence of H-bonded Br–, those of other Laminaria thallus parts (apart from the blade) in Fig. 3, viz. the methanol extract enriched in polyphenols, the cortical tissue, the stipe, and the meristem, appeared to contain bromide in the position of a Cl defect in a KCl crystal. It is possible to interpret this result as due to the presence of free Br– which co-precipitates with KCl upon cooling of the sample for the XAS measurement, but the concept that Br– is also associated with solid KCl under physiological conditions in Laminaria cannot be excluded. It should be noted that the highest concentrations of K+ and Cl– ions in the cortex and medulla of dry weight Laminaria (120,000 and 109,000 ppm, respectively (Verhaeghe ), which turn out to be equimolar amounts upon correction for the atomic weight) imply that 22.9 % of the solid material was KCl. The new data are in accordance with the report of bromine localization in Laminaria (Verhaeghe ), which suggests that bromine is less diffusible than iodine in stipe tissues and is proportionally more abundant in strong association with the cell walls. By merging these different approaches – microchemical imaging (Verhaeghe ), XAS, ICP-MS, EPR, and whole-blood antioxidant assays – this study sought to provide new hypotheses about bromination processes and the function of bromine in brown algal kelps.

Fig. 1.

Laminaria tissues versus NaBr solution and seawater. Br K-edge experimental (grey line) and simulated (black line) EXAFS (left) and phase-corrected Fourier-transforms (right): (A) NaBr solution (number and type of atoms @ distance in Å (Debye-Waller factor as 2σ2 in Å2): 9.5 O @ 3.335 (0.030)), (B) seawater (9.5 O @ 3.300 (0.028)), (C) L. digitata stressed with oligo-GG (5.0 O @ 3.262 (0.039)), (D) fresh L. digitata blotted (1.8 O @ 3.254 (0.023)), (E) lyophilized L. digitata (2.2 H @ 1.283 (0.002) and 2.1 O @ 3.238 (0.006)), and (F) L. digitata holdfast (1.7 O @ 3.443 (0.009)). See also Supplementary Table S1.

Fig. 3.

Different Laminaria extracts and tissues. Br K-edge experimental (grey line) and simulated (black line) EXAFS (left) and phase-corrected Fourier-transforms (right): (A) polyphenol-enriched extract (number and type of atoms @ distance in Å (Debye-Waller factor as 2σ2 in Å2): 1.4 K @ 3.173 (0.008)), (B) cortical tissue (1.8 K @ 3.203 (0.010) and 0.5 C @ 1.907 (0.008)), (C) stipe (2.0 K @ 3.211 (0.007) and 0.1 C @ 1.864 (0.008)), and (D) meristem (2.4 K @ 3.207 (0.010) and 1.890 (0.007)). See also Supplementary Table S1.

Fig. 2.

Brown algal haloperoxidases and brominated amino acids. Br K-edge experimental (grey line) and simulated (black line) EXAFS (left) and phase-corrected Fourier-transforms (right): (A) 3,5-dibromotyrosine/BN (number and type of atoms @ distance in Å (Debye-Waller factor as 2σ2 in Å2): 1.0 diBrTyr with C @ 1.882 (0.003)), (B) A. nodosum bromoperoxidase (1.0 diBrTyr with C @ 1.906 (0.002)), (C) 4-bromophenylalanine/BN (1.0 phenyl with C @ 1.856 (0.004)), (D) L. digitata iodoperoxidase (1.0 indolyl with C @ 1.847 (0.003)), and (E) aqueous 5-bromotryptophan (1.0 indolyl with C @ 1.862 (0.006)). See also Supplementary Table S1.

A comparison of the pattern of peaks in the Fourier transform of the Br EXAFS for the native L. digitata VHPOs (Fig. 2D for iodoperoxidase, Supplementary Fig. S3C for bromoperoxidase) with a number of brominated amino acids (3-Br-tryptophan, 4-Br-phenylalanine, and 3,5-di-Br-tyrosine in Figs. 2E, 2C, and 2A, respectively) showed evidence for incorporation of Br into the aromatic moiety of the amino acids. This was, however, not in the form of 3,5-dibromo-tyrosine as established earlier for the bromoperoxidase of Ascophyllum nodosum by XAS (also included in Fig. 2 as trace B) and mass spectrometry (Feiters ). Simulations of the Laminaria haloperoxidase Br XAS with structural models based on 5-Br-tryptophan or 4-bromo-phenylalanine, respectively, confirmed that the Br was attached to an aromatic group, but did not allow unambiguous identification of the type of amino acid (3-bromo-tyrosine, 4-bromophenylalanine, 5-bromotryptophan), except for the exclusion of 3,5-dibromo-tyrosine. The native bromoperoxidase of A. nodosum contained Br in the surface tyrosine residues 398 (rarely monobrominated) and 447 (frequently dibrominated) (Feiters ). In fact, the alignment of the different VHPO proteins did not show a strict conservation of tyrosine residues at these two positions, but phenylalanine residues in L. digitata VHPOs (Supplementary Fig. S5). Both types of native VHPOs from Laminaria were brominated; the enzymes appeared to have been exposed to conditions in the algae or during enzyme extraction where they were brominated. It should be noted that the bromoperoxidase is capable of brominating itself, whereas the iodoperoxidase is unable to react with bromide (Colin ).

XAS of iodine in Laminaria (Supplementary Fig. S4) shows that it was present as iodide, which is hydrogen-bonded to large biomolecules in fresh Laminaria, but partly mobilized as hydrated iodide upon addition of an oligo-GG elicitor, as reported earlier (Küpper ), and of H2O2. The iodine EXAFS showed no evidence of incorporation in KCl crystals, as found for the bromine. Iodide has been found (Verhaeghe ) to be associated more exclusively than bromide with peripheral tissue, where the potassium and chloride concentrations are relatively low; moreover, the iodide ion would fit even more poorly in the KCl lattice than the bromide ion, due to its larger size.

Consistent with previously published results (Carpenter ; Palmer ; Thomas ), Laminaria emitted predominantly brominated halocarbons in the unstressed steady state, while an oxidative burst following GG treatment resulted in strongly increased emissions of iodocarbons (Supplementary Fig. S6). Arguably, the more physiological form of oxidative stress (caused by GG) was met by a mobilization of the iodide antioxidant reservoir, thus providing cell-wall-bound haloperoxidase enzymes both with their preferred substrate and also at a locally higher concentration than bromide, leading to an increased incorporation of iodine into organic molecules, resulting in mono- and multiple iodinations. However, it has to be noted that the total emission rate of iodine via iodocarbons was at least three orders of magnitude lower than the inorganic iodine efflux previously reported (Küpper ; Chance ). This efflux was dominated by CH2I2, CH3I, CHBr2I, and C2H5I (Supplementary Fig. S6B). Interestingly, both vanadium iodoperoxidase and vanadium bromoperoxidase genes are upregulated after GG elicitation (Cosse ) and in silico protein sequence targeting suggested cell-wall localization (Colin , 2005). Collén found that adding H2O2 to Meristiella gelidium increased production of bromo- and chlorocarbons whereas iodocarbon levels were unaltered. Comparison of exogenous oxidative stress (exogenous H2O2) versus a GG-elicited oxidative burst (Supplementary Fig. S6A,B) suggested that in L. digitata, simple extracellular addition of H2O2 did not adequately mimic the rapid physiological ROS formation plus mobilization of accumulated iodide and volatile halocarbons that occurs during a GG-triggered oxidative burst. Under stress conditions, spontaneous chemical reaction of I– with H2O2 and the activity of constitutive algal haloperoxidases are likely to generate hypoiodous acid (HOI) in equilibrium with I2. HOI and I2 are highly reactive and readily react with dissolved organic matter to produce organoiodine compounds (Truesdale ; Bichsel and von Gunten, 1999; Carpenter ), analogous to bromocarbon synthesis by HOBr/bromoperoxidase (Theiler ). The rates of production of iodine- and bromine-containing compounds will depend on the relative concentrations of I– and Br–, the specificity of the haloperoxidase enzymes, dissolved organic matter availability, and type. Carbon skeletons could potentially originate from the alga itself (e.g. from the release of free fatty acids), which would then be iodinated and oxidized in the strongly oxidizing conditions prevalent during the oxidative burst (Küpper ). Critically for the production of volatile halocarbons, the substantial I– efflux would shift the balance between Br– and I–, increase HOI production relative to HOBr, and swamp the haloperoxidase system with iodide such that iodocarbons dominate the blend of halocarbon volatiles produced. The absence of a bromide efflux during oxidative stress (supporting information) contrasted with the strong iodide efflux observed under the same conditions (Küpper ; Chance ) due to the predominance of iodide accumulation compared to bromide.

An antioxidant assay based on phorbol myristate acetate-stimulated human blood cells showed that KBr has significant antioxidant activity only at the highest concentration tested, 100mM (Fig. 4). At all lower concentrations (0.01, 0.1, 1 and 10mM), no significant anti- or pro-oxidant effect was detected. Remarkably, in the same assay, iodide showed a pro-oxidant effect at 0.01, 0.1, 1, and 10mM which, however, was attenuated at 10mM KI by the presence of 1mM KBr. Only the highest iodide concentration, 100mM, had a strong antioxidant effect in this assay.

Fig. 4.

Neutrophil antioxidant assay at varying concentrations of bromide and iodide.

The results of the in vitro EPR assay that discriminated between scavenging of superoxide and hydroxyl radicals are shown in Fig 5. KBr showed a trend towards scavenging of superoxide at concentrations >1mM, reaching statistical significance at 100mM. KI failed to scavenge superoxide at the same concentrations and showed a trend for pro-oxidant activity. In contrast, both KBr and KI exhibited complex interactions with hydroxyl radicals, with a pro-oxidant effect at moderate concentrations (10–100 μM) being overwhelmed by antioxidant effects at high concentrations (significant at 100mM). The results show that the reactivity of the inorganic halide with radicals is complex and concentration dependent. High Br– concentrations (>1mM) were found to scavenge superoxide whilst equivalent concentrations of I– tended to exacerbate radical generation. In contrast, both halides increased radical generation at moderate (10–100 μM) concentrations and were equally effective at sequestering hydroxyl radicals at concentrations >1mM. Indeed, similar concentration-dependent pro- and antioxidant behaviour is known for a range of biological antioxidants (Long and Halliwell, 2001). At present, there is no evidence to support a mechanism for the phenomenon in the case of iodide and bromide, but it is tempting to hypothesize that it might (as is the case with some other antioxidants) hinge around the balance between reduction of metal ions to more reactive reduced forms for mediating Fenton chemistry (e.g. Fe2+; pro-oxidant) and radical-scavenging activity (antioxidant).

Fig. 5.

Electron paramagnetic resonance determination of ROS by spin trapping. Effect of increasing concentrations of KBr (A, B) and KI (C, D) on spin adduct formation in the presence of superoxide (pyrogallol; A, C) and hydroxyl (menadione; B, D) radical generators. *P < 0.05, **P < 0.01, ***P < 0.001 compared to control (0mM KI or KBr): Dunnett’s post-test after one-way ANOVA.

This study shows that L. digitata sporophytes contain bromine as bromide, similarly to the more pronounced accumulation of iodide, but to a lesser degree and also bound to aromatic molecules. The chemical in vivo speciation of bromide in Laminaria resembles that of iodine – it is overwhelmingly associated with biomolecules via hydrogen bonds. Also similarly to the previously documented case of iodide (Küpper ), oxidative stress results in increased hydration. However, and even though in principle the same reaction pathways would apply for both bromide and iodide (such as halide-assisted disproportionation of hydrogen peroxide; Vilter, 1995; Küpper ), bromide appears to be less suitable as an antioxidant than iodide, based on both thermodynamic and kinetic considerations regarding the reactions of halides with the oxidants relevant in a living system. Its function is likely not only that of an antioxidant but also, as found in other brown algae not showing a specific iodine concentration, a matter of interacting macromolecules (such as cross-linking in Fucales; Potin and Leblanc, 2006), and involvement in different and complementary roles for chemical defence, as also found in red algae (such as disruption of the quorum-sensing systems of bacterial biofilms; Butler and Carter-Franklin, 2004).

However, the antioxidant function of iodide clearly depends on the nature of the oxidant considered, as established by neutrophil oxidative burst and EPR assays. With regard to superoxide and as highlighted by both the neutrophil oxidative burst and EPR assays, iodide was actually pro-oxidant between 0.1–10mM; only higher concentrations were antioxidant. However, superoxide did not accumulate during oxidative stress in Laminaria but was rapidly converted to hydrogen peroxide by the action of superoxide dismutase (Küpper ) and scavenged by bromide (this study; Fig. 5). In contrast, iodide is a very efficient antioxidant against hydrogen peroxide, hydroxyl radicals, and ozone (Küpper ), and it is well established that at least hydrogen peroxide (Küpper , 2002) and ozone (Palmer ; Küpper ) can exist at concentrations physiologically relevant for Laminaria. In one-electron oxidations, only iodide – not bromide – reacts exergonically with hydroxyl and superoxide radicals (Luther III, 2011). Among the two-electron reactions, iodide is still by far the most favourable reaction partner for many biologically relevant oxidants (ozone, singlet oxygen, and hydrogen peroxide), even though bromide and chloride show favourable (albeit considerably smaller) ΔGR values (Luther III, 2011) for the reaction with ozone and hydrogen peroxide (and the EPR data suggest that bromide might have some additional benefits over iodide with regard to superoxide radical). Similarly, the reaction of iodide with ozone, singlet oxygen, and hydrogen peroxide was several orders of magnitude faster than that of bromide or chloride (Table 2). Finally, it should be highlighted that iodide is locally the predominant halide species and could be the preferred substrate of cell-wall-localized vanadium haloperoxidases over bromide and chloride (Verhaeghe ). Haloperoxidases also considerably accelerate the halide-assisted disproportionation of hydrogen peroxide (Vilter, 1995; Küpper ).

Table 2.

Kinetics of oxygen species with halides and other reductants.

Compound	k12 (M–1 s–1)	Source and notes
O3 reactions with:
I–	1.2×109	Liu et al. (2001)
Br–	2.48×102	Liu et al. (2001)
Cl–	<3×10–3	Hoigné et al. (1985)
Ascorbate	4.8×107	Kanofsky and Sima (1995)
Glutathione	2.5×106	Kanofsky and Sima (1995)
Singlet oxygen (1O2) reactions with:
I–	1×108	Rosenthal (1976; aprotic solvents)
	8.7×105	Wilkinson et al. (1995, p 896; pH ~7)
Br–	1.0×103	Wilkinson et al. (1995, p 895; in D2O)
Cl–	1.0×103	Wilkinson et al. (1995, p 895; in D2O)
Ascorbate	8.3×106	Wilkinson et al. (1995, p 904; pH 6.8)
Glutathione	2.4×106	Wilkinson et al. (1995, p 883; in D2O, 310 K, pD 7.4)
OH radical (⋅OH) reactions with:
I–	1.2×1010	Buxton et al. (1988, p 527)
Ascorbate	1.1×1010	Buxton et al. (1988, p 700)
Glutathione	1.3×1010	Buxton et al. (1988, p 723; pH 5.5)
Dimethyl sulphoniopropionate	3×108	Sunda et al. (2002)
Dimethyl sulphide	1.9×1010	Buxton et al. (1988)
Dimethyl sulphoxide	6.6×109	Buxton et al. (1988)
Superoxide (O2 –) reactions with:
I3 –	1×108	Bielski et al. (1985, p 1063; no data available for I–)
Ascorbate	2.7×105	Bielski et al. (1985, p 1069; pH 7.4)
Glutathione	2.4×105	Bielski et al. (1985, p 1075; pH 7.8)
Hydrogen peroxide (H2O2) reactions with:
I–	0.69	Mohammed and Liebhafsky (1934)
Br–	2.3×10–5	Mohammed and Liebhafsky (1934)
Cl–	1.1×10–7	Mohammed and Liebhafsky (1934)
Ascorbate	2	Polle and Junkermann (1996)
Glutathione	2–20	D’Autréaux and Toledano (2007)
Glutathione peroxidase	6×107	Flohe et al. (1972)

The EPR results highlight the importance of having the correct blend of antioxidants present in the correct compartment in order to ensure beneficial as opposed to detrimental or neutral effects. For example, high concentrations of Br–, but not I–, could represent a useful antioxidant in defence against superoxide (e.g. oxidative burst), whilst high (but not moderate) concentrations of both Br– and I– could protect against hydroxyl radical (e.g. derived from hydrogen peroxide via Fenton chemistry). From an evolutionary perspective, therefore, it would prove advantageous to maintain large stores of halides in inactive (caged) forms (i.e. bound to organic entities) and to release them in high concentrations in response to stress. A combined assault with Br– and I– would have the added benefit of targeting different ROS species, thus providing comprehensive protection.

Overall, in Laminaria, bromide compares with iodide as follows. (i) There is no strong accumulation of bromide from seawater, and no detectable bromide efflux upon oxidative stress. (ii) An oxidative burst results in a shift to increased iodocarbon emissions (this study; Palmer ; Thomas ). (iii) The reaction of bromide with most of the biologically relevant oxidants is not favourable, neither in a thermodynamic nor in a kinetic sense. Therefore, the results suggest that, while the extracellular antioxidant defence of Laminaria is primarily based on iodide, iodide and bromide may complement each other in the face of different oxidants such as superoxide and that the role of bromine and brominated compounds in macroalgae is more complex than previously thought (Table 3).

Table 3.

Comparison of the key features of iodine and bromine metabolism in Laminaria.

Feature	Iodine	Bromine
Accumulation factor (from seawater to Laminaria)	104–105	1–101
Efflux upon oxidative stress	Yes	Not detected
Halocarbon emission	Iodinated halocarbons are emitted mainly after the oxidative burst	Bromocarbons are emitted at high rates under unstressed, steady-state conditions
Detection of oxidized species upon oxidative stress	No	No
Antioxidant effect in whole-blood assay	Yes	Partial

This work constitutes the first study about the physiology of bromide as a potentially comparable antioxidant. The results show that bromide complements iodide as an inorganic, extracellular antioxidant in Laminaria, in particular against superoxide, but also that it is less effective against most other biologically relevant oxidants and that its function is more diverse than that of iodide. It is proposed to call the preferential accumulation and targeted release of iodide upon oxidative stress, in comparison to bromide, the ‘iodide switch’. This work suggests that the accumulation of iodide and the iodide switch are central and unique features of the evolution of morphologically complex, large brown algal kelps such as Laminaria. The evolutionary origin of this unique antioxidant system remains one of the most intriguing research questions in this context.

Supplementary data

Supplementary data are available at JXB online.

Supplementary Fig. S1. Scheme of Laminaria digitata showing the different parts of the thallus

Supplementary Fig. S2. Br K-edge XANES and EXAFS of four representative samples: seawater and L. digitata stipe, cell wall, and lyophilized L. digitata rehydrated with H2O2

Supplementary Fig. S3. Br K-edge EXAFS and phase-corrected Fourier-transforms of lyophilized L. digitata rehydrated with H2O2, cell wall, L. digitata native bromoperoxidase, and inactive Escherichia coli-expressed iodoperoxidase with added Br

Supplementary Fig. S4. I K-edge EXAFS and phase-corrected Fourier-transforms of fresh L. digitata, L. digitata stressed with oligo-GG or H2O2, and lyophilized and rehydrated L. digitata

Supplementary Fig. S5. Partial amino-acid alignment of vanadate-dependent haloperoxidases

Supplementary Fig. S6. Release of volatile low-molecular-weight halocarbons from L. digitata plantlets

Supplementary Table S1. Parameters resulting from iterative refinement of simulations of the Laminaria bromine-edge EXAFS

Supplementary Table S2. Parameters resulting from iterative refinement of simulations of the Laminaria iodine-edge EXAFS